biocontext-knowledge_skill

Overview

This skill provides programmatic access to integrated biomedical knowledge for gene and protein annotation. It queries UniProt, AlphaFold, STRING, Reactome, GO, PanglaoDB, PubMed/Europe PMC, OpenTargets and other resources through ov.biocontext. Use it to add biological context to omics results, validate markers, explore pathways, or investigate drug–disease links.

How this skill works

The skill calls BioContext tools via ov.biocontext to retrieve structured records from 49 biomedical databases. Functions accept common identifiers (gene symbol, UniProt accession, Ensembl ID) and species codes, returning JSON or DataFrame-ready results like protein domains, GO terms, pathway membership, marker scores, publications, and clinical links. You can list available tools or call any tool by name for custom GraphQL or REST queries.

When to use it

• Annotating differential expression or cluster marker lists with protein function and pathways.
• Validating candidate cell-type markers against PanglaoDB marker sets.
• Finding protein interactions and network partners via STRING before network analysis.
• Searching literature and preprints for gene-specific evidence and recent findings.
• Exploring drug targets, associated diseases, and clinical trials for translational follow-up.

Best practices

• Always set the correct species code (human default is 9606 or homo_sapiens) to avoid empty results.
• Batch queries with small delays to reduce rate limiting and avoid timeouts when annotating large gene lists.
• Use multiple identifier types (gene symbol, UniProt ID, Ensembl ID) if one returns no hits.
• Limit GO and pathway sizes when inspecting many genes to keep responses lightweight.
• For OpenTargets use valid GraphQL queries and test them on a single gene before batch execution.

Example use cases

• Annotate top 50 DEGs with UniProt function, Reactome pathways, and top GO terms for reporting.
• Confirm that scRNA-seq cluster markers match PanglaoDB T-cell markers prior to cell-type naming.
• Retrieve AlphaFold structure confidence for candidate proteins before designing domain-focused experiments.
• Search Europe PMC for recent papers linking a gene to a phenotype or assay used in your study.
• Query OpenTargets to list diseases and drugs associated with a gene and check active clinical trials.